CosmoCoffee

I've added a parameter to the standard CosmoMC parametrization and when they converge, I'm getting flat contours where my new parameters are completely independent of the standard cosmology ones and then much narrower than the priors I set?

Any reason why my parameters could be marginalising to such a narrow final boundary?

Thanks for any help

Hi Luke! Could you please give some more details? E.g. what these parameters are (are they recombination parameters you are playing with?), what prior did you originally set, which datasets are you using, what level of convergence are you requiring?

Cheers,
Sunny

Update: Found! It seems to have originated from the proposal distribution so I tweaked that a little bit and now it seems to be giving actual covariances rather than just remaining degenerate to the code.